Your chats with Claude just became training data – but there’s a way to opt out
Anthropic’s Claude AI Chatbot to Utilize User Transcripts for Training, but Users Can Opt Out
In a recent policy change announced by Anthropic, the company behind the popular AI chatbot Claude, users will soon have their chats and interactions logged and used for training purposes. This move has sparked debate within the tech community, with some expressing concerns about data privacy and others seeing it as a necessary step towards improving AI performance. The updated policy allows users to opt out of this feature until September 28th.
According to Anthropic’s statement, the logs of user interactions will be used for training models, model improvement, and strengthening safety guardrails. Users will see a notification alert about the data usage policy until September 28th, which will inform them that by participating in the program, they can help improve model safety and accuracy in detecting harmful content.
However, users are not obligated to participate, as noted in Anthropic’s clarification of its new user data policy. Existing Claude users can choose to disable the "You can help improve Claude" toggle by navigating to Settings > Privacy > Help Improve Claude after September 28th.
The updated training policy only applies to new chats with Claude or those that are resumed, but excludes old conversations. The company also outlined additional changes to its data retention rules, stating that user data will be retained for up to five years, unless it is deleted manually by the user. Chats that have been manually deleted will not be used for AI training.
This move can be seen as a response to the industry’s growing concern about data scarcity and the need for high-quality training datasets. Anthropic’s decision has sparked discussion within the tech community, with companies like Apple using Claude as a potential candidate to enhance Siri’s capabilities.
The updated policy affects users across all Claude plans, including Free, Pro, and Max. However, it does not apply to specific use cases such as work environments or when connected to third-party platforms.
One of the key concerns associated with this change is the possibility that sensitive information could be inadvertently collected during user interactions. To mitigate this risk, Anthropic emphasizes the importance of user data privacy and has outlined various steps users can take to protect their information.
In related news, there have been reports of sophisticated AI tools being used for scams and financial gain. One recent case involved a real estate firm owner who came across a suspicious phone number while looking up Royal Caribbean’s customer service on Google. The phone number found in the top of the Search page was not genuine but rather a scammer-operated hotline.
AI overviews and AI Mode results can inadvertently direct users into potentially hazardous situations. In such cases, experts have warned that users must exercise caution when dealing with AI-enhanced search features.
Another area where OpenAI has announced significant advancements is with the release of GPT-5. This large language model boasts enhanced capabilities compared to its predecessor but still lacks tangible evidence demonstrating the validity of its features.
As we navigate an increasingly interconnected digital landscape, it becomes more evident that human skills and abilities will need to be adapted in line with the ever-evolving AI environment.
AI Advancements and Associated Risks
Anthropic’s move towards utilizing user data raises questions about accountability within the AI industry. One notable initiative aiming to address this issue is parent control for large language models, a development expected to positively impact safety standards across multiple platforms.
As concerns regarding user protection continue to rise, researchers are now urging policymakers to take an active role in defining regulations aimed at curtailing misuse of AI tools and preventing them from falling into the wrong hands.
